190 ME. F. J. COLE ON THE STRUCTURE AND MORPHOLOGY OF 



canals." Before this can take place, however, the anlage of the system must be formed, 

 and Allis finds (ji. 535) tliat " each group [/. e. line of sense organs] develops from the cord 

 of cells lying iix the deeper layers of the eiiidermis, and each cord from a special sensory 

 thickening, which when first seen from the surface in S2:)ecimens liardened in chromic or 

 picro-sulphuric acid, appears as a large, whitish, and slightly raised spot." From this 

 anlage all the canal organs arise, and " in the early stages of their development lie 

 below the surface, but they soon push through the overlying epidermal cells, and their 

 upper central portions become exposed. Each pit organ subsequently sinks slightly 

 below the surface, and a little epidermal pit is formed above its central portion " (p. 536). 

 The canal organs are develojied as above described. Combining Allis's results first 

 with the work of Balfour and the eai'lier embryologists and then with the recent 

 observations of H. V. Wilson, Mitrophanow, and Locy, we are enabled to draw up the 

 following scheme * of the development of tlie lateral canal system, which is known to 

 apply to some, if not to all, fishes : — (1) A small sensory anlage arises in the neighbourhood 

 of tlie ultimate position of the auditory organ, the differentiation of which gives rise both 

 to the auditory organ and the system of sensory canals. (2) This anlage grows forwards 

 and backwards as a non-segmental cord of cells. In front this cord bifurcates and forms 

 the anlagen of the supra- and infra-orbital canals. Behind it forms the anlage of the 

 lateral or body canal. (3) Along the non-segmental cord arise series of sense organs, 

 which, in the case of the body canal are segmental, but are not known to be such in 

 the other canals. (4) Each sense organ sinks down, carrying \he skin with it, so that it 

 ixltimately lies at the bottom of a groove. The lips of the groove fuse — each sense 

 organ therefore lying in a short tube opening at each end on to the surface by a pore 

 ( = "half pore," Allis). (5) The short tube extends both ways and fuses with the two 

 adjacent tubes, adjacent half pores fusing to form a single " primary " pore which in most 

 fishes forms the external aperture of a dermal tubule. In the Ganoid fishes the primary 

 pore, by a process of repeated dichotomy, produces what is known as a " dendritic system." 

 A lateral canal is therefore formed by the end-to-end fusion of pieces or segments, and 

 not by the formation of a primitive continuous furrow, and it further follows that there 

 must have been primitively a dermal tubule between every two adjacent sense organs. 



Bateson (1889-90, lo) fails to distinguish between terminal buds and pit organs, and 

 although pointing out that these two sets of organs have not the same histological 

 structure, seems to have overlooked the previously described fact that the pit organs 

 belong to the lateral line system. According to Coggi (1891, 42), Savi's vesicles 

 develop in the same Avay in Torpedo as the sense organs of the lateral canals, but his 

 description differs somewhat from Balfour's observations on Scyllinm. Ayers (1892, 7), 

 who occupies an isolated position on some aspects of the question, believes Savi's 

 vesicles to have been produced retrogressively from canal organs, and considers that 

 they are " without doubt descendants of the canal organs " (p. 164). He bases this 

 opinion on the observations of Fritsch, which he thinks show that the Savian vesicles 



* This scheme ignores the opereulo-mandibular caual, of the development of which our knowledge is still very 

 imperfect. I sec uo roa-^on, however, wh}' it should not develop in just the same way as, for example, the infra- 

 orbital does, i. e. by branching from the main trunk. 



